Date: Mon, 11 May 2015 09:53:15 -0400 From: Ed Maste <emaste@freebsd.org> To: Mark Johnston <markj@freebsd.org> Cc: Julian Elischer <julian@freebsd.org>, "src-committers@freebsd.org" <src-committers@freebsd.org>, "svn-src-all@freebsd.org" <svn-src-all@freebsd.org>, "svn-src-head@freebsd.org" <svn-src-head@freebsd.org> Subject: Re: svn commit: r282744 - in head/sys/cddl/dev/dtrace: amd64 i386 Message-ID: <CAPyFy2BkgbHnXhD8zPvopKnE1RL7MKNR2i6q%2Br4rdnNBEvUK0w@mail.gmail.com> In-Reply-To: <20150511024754.GA65659@raichu> References: <201505102227.t4AMRn1g007294@svn.freebsd.org> <55500C0D.7090802@freebsd.org> <20150511024754.GA65659@raichu>
next in thread | previous in thread | raw e-mail | index | archive | help
On 10 May 2015 at 22:47, Mark Johnston <markj@freebsd.org> wrote: > > I've caught several cases (specifically in merging DTrace code) where the > upstream code block is so large the merger didn't even notice the > #ifndef __FreeBSD__ was present and thought that the merge was fine because > the upstream diff applied cleanly. Indeed - leaving large blocks of #if'd out upstream code can easily lead to a false sense that everything was successful in a merge. It's actually better that the merge creates a conflict: the merger then needs to understand the change and determine if it actually applies to FreeBSD or not, rather than just accepting the patch. Both approaches have arguments for and against, but ultimately I think the decision rests with the one doing the work.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAPyFy2BkgbHnXhD8zPvopKnE1RL7MKNR2i6q%2Br4rdnNBEvUK0w>